Revised 19 October 2026 · v1.0.0 · what changed

Code: github.com/herrrickshaw/FCI-warehouse-MSPvsMSV

Watch the walkthrough

Watch on YouTube: youtu.be/KdOY3vWwzuM

Why it is useful

• Agricultural-policy analysts, journalists and students asking 'does MSP actually hold in the market?' get a ready table of mandi modal prices against MSP, crop by crop, with sources recorded in a manifest.
• It replaces stitching together CACP, Agmarknet, DFPD bulletins and FCI data by hand. The data/ folder holds 49 CSVs, and data/MANIFEST.md says where each came from and what is missing.
• Evidence from the run: the mandi table shows 92% of bajra and 84% of moong market reports below MSP on 23 July 2026, against 19% for wheat. Two models also ran unchanged in a fresh environment.
• The holding-economics model produced a clear result: for every commodity tested, estimated storage and financing cost exceeded price appreciation over 12, 24 and 36 months, which bears directly on the 'why not buy and hold' question behind MSV.

What it is for

FCI-warehouse-MSPvsMSV evaluates whether a Minimum Support Volume guarantee could extend effective price support beyond rice and wheat. It contains cited reports (report/), CSV data with a provenance manifest (data/), a self-contained HTML dashboard (dashboard/), a Colab notebook, and several Python models: holding economics, OMSS grain-processing unit economics, rice-bran oil savings, and a real farm income index.

This post uses the MSP series, the mandi-vs-MSP table and the two models that ran cleanly.

Who it is for

• Agricultural-economics students and researchers
• Journalists covering procurement, MSP and food stocks
• Policy teams comparing price-support designs
• Developers wanting a worked, sourced Indian agri dataset

What you need

Get the code and an AI assistant

You do not have to be a programmer. Get the repository, give it to an AI assistant, and ask in plain language. Any of these routes works:

git clone https://github.com/herrrickshaw/FCI-warehouse-MSPvsMSV.git
cd FCI-warehouse-MSPvsMSV
claude      # or: gemini

Install

1. Clone the repository

   git clone https://github.com/herrrickshaw/FCI-warehouse-MSPvsMSV.git
   cd FCI-warehouse-MSPvsMSV

2. Create an environment and install dependencies

   python3 -m venv .venv && source .venv/bin/activate
   pip install pandas numpy

Quick start: commands and what you should see

View the MSP vs mandi table

cat data/mandi_vs_msp_20260723.csv

16 crops with MSP, number of market reports, median modal price, share of reports below MSP and median gap.

Run the grain-processing unit-economics model

cd data && python3 omss_processing_model.py

EBITDA per quintal for wheat-to-atta and paddy-to-rice, plus a 300 TPD flour-mill scale case.

Run the holding-economics model

cd data && python3 holding_economics.py

Summary, horizon-return and post-crash-episode tables per commodity; it rewrites three CSVs next to the script, so run it in a copy if you want the repo untouched.

Sample outputs

Real output from a run on the date shown in the sources line; trimmed to fit.

MSP vs mandi sample table (data/mandi_vs_msp_20260723.csv)

crop     MSP 2025-26  market reports  modal median  % reports below MSP  median gap vs MSP
wheat        2425           768         2500.0          19                 +3%
paddy        2369           335         2385.0          36                 +1%
gram         5650           231         5900.0          22                 +4%
maize        2400           226         2200.0          75                 -8%
soybean      5328           195         7050.0           2                +32%
moong        8768           161         6901.0          84                -21%
tur          8000           129         7531.0          70                 -6%
bajra        2775            84         2382.0          92                -14%
ragi         4886            14         3550.0          93                -27%

Prices in Rs per quintal. Trimmed from 16 crops. Built from the 23 July 2026 Agmarknet daily feed (17,247 records) per MANIFEST.md. Crops with few reports (ragi 14, sunflower 2) are thin evidence.

Holding economics (14-year window, % per year)

commodity   price CAGR  carry cost  net annual  verdict
paddy             4.84       20.33      -15.48  CARRY EXCEEDS APPRECIATION
wheat             5.19       16.99      -11.80  CARRY EXCEEDS APPRECIATION
maize             3.81       18.50      -14.69  CARRY EXCEEDS APPRECIATION
arhar_tur         5.74       13.67       -7.93  CARRY EXCEEDS APPRECIATION
sugar             2.33       16.59      -14.26  CARRY EXCEEDS APPRECIATION

Rows trimmed; all 16 commodities show the same verdict. Carry costs combine storage tariffs and financing assumptions that are cited in the repo's report, not in this output.

Grain-processing unit economics

python3 omss_processing_model.py
WHEAT->ATTA OMSS-sourced: grain 2550 | revenue 2926 | gross 376 | EBITDA 196/qtl (6.7% margin)
WHEAT->ATTA mandi-sourced: grain 2500 | revenue 2926 | gross 426 | EBITDA 246/qtl (8.4% margin)
PADDY->RICE open-market: paddy 2385 | revenue 2520 | EBITDA -15/qtl (-0.6% margin)
SCALE: 300 TPD flour mill | EBITDA Rs 17.6 cr | capital employed ~Rs 80 cr | ROCE ~22%

Deterministic model with labelled input assumptions, not a forecast.

Ways to use it

Sample prompts for your AI assistant

Copy, paste and edit the parts in [brackets].

• Read data/MANIFEST.md and list which datasets cover procurement, stocks, storage and prices, with their known gaps.
• Using data/msp_series.csv, compute the MSP CAGR for wheat and paddy since 2010-11 and chart both.
• From mandi_vs_msp_20260723.csv, which crops have more than half of market reports below MSP? Why might the sample sizes matter?
• Explain the assumptions in holding_economics.py and which one the verdict is most sensitive to.
• Summarise the case for and against a Minimum Support Volume from report/msp-vs-msv-main-report.md in 200 words.

Limits and honest caveats

Sources: the FCI-warehouse-MSPvsMSV README, data/MANIFEST.md, data/msp_series.csv, data/mandi_vs_msp_20260723.csv and the Python models; omss_processing_model.py and holding_economics.py run on 19 October 2026 on a copy of the data folder (latest commit 3 October 2026).

About this article: Researched, written and edited by Umashankar Triplicane Dwarakanathan, with AI research assistance; every figure is meant to trace to the primary source cited. See the Editorial Policy for how sourcing, AI use and corrections work.